In an audio data processing process, interframe distortion often occurs because transition between audio frames is not sufficiently smooth. For example, a popping noise occurs between two audio frames. This phenomenon is caused because of different processing algorithms of two adjacent audio frames.
An existing basic method for resolving the foregoing problem is performing smoothing processing on a distorted part. For example, smoothing processing is performed on data corresponding to two different audio frames of a processing algorithm by using a low-pass filter. Specifically, a distortion location is determined; data corresponding to an audio frame before the processing algorithm is switched and data corresponding to a current audio frame are separately obtained; low-pass filtering processing is separately performed on data corresponding to an audio frame before the processing algorithm is switched and data corresponding to a current audio frame, and the processed data is placed to an original position, so as to smooth the distorted part. However, because in the foregoing method, data corresponding to an audio frame before switching is obtained, and data corresponding to a current audio frame is obtained, only off-line processing is performed. That is, entire audio data is processed after the audio data is completely recorded, and on-line processing, that is, processing while playing, cannot be performed. That is, the off-line processing manner cannot resolve interframe distortion caused because of switching of the processing algorithm in a playing process.